Childhood adversity and multi-dimensional cognitive dysfunction in major depressive disorder: investigating heterogeneous effects across gender and educational in the UK Biobank

Abstract

Background

Major Depressive Disorder (MDD) is characterized by cognitive dysfunction that persists as one of the most profoundly disabling mental health conditions worldwide. Childhood adversity (CA) is not only a significant risk factor for the onset and course of MDD, but it is also strongly linked to decreased cognitive performance in MDD. Existing studies feature small sample sizes, single evaluation indicators, inconsistent outcomes, and few studies on the effects of demographic variables.

Method

This study used cross-sectional data (N = 2353) from the UK Biobank to investigate the relationship between childhood adversity (CA) and multi-dimensional cognitive function in patients with major depressive disorder (MDD), as well as the moderating effects of demographic variables. A generalized linear model was used to assess the relationships between five types of CA (physical neglect, emotional neglect, sexual abuse, physical abuse, and emotional abuse), as well as cumulative adversity scores, and multi-dimensional cognitive tests (fluid intelligence, reaction time, and pairs matching test).

Result

The findings revealed that higher scores of physical abuse (β=-0.420, 95% CI=-0.729 to 0.111, P = 0.008), physical neglect (β=-0.578, 95% CI=-0.922 to -0.235, P = 0.001) and cumulative adversity (β=-0.169, 95% CI=-0.272 to -0.066, P = 0.001) were substantially associated with poorer fluid intelligence, whereas emotional abuse (β = 0.285, 95% CI = 0.020 to 0.549, P = 0.035) and physical neglect (β = 0.328, 95% CI = 0.020 to 0.636, P = 0.037) were associated with poorer performance on the pairs matching test. The stratified analysis demonstrated that the adverse impact of childhood adversity on cognitive function was exacerbated among female patients, while there was no discernible correlation in male patients. Highly educated patients showed poorer fluid intelligence while less educated patients showed poorer pair matching.

Conclusion

The study reveals that the influence of childhood adversity on cognition is not uniform but is instead shaped by key sociodemographic moderators, including gender and education. This suggests that its effects are driven by an interplay of neurobiological mechanisms and divergent adaptive cognitive strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-026-04029-z.

Introduction

Major Depressive Disorder (MDD) is one of the mental illnesses with the highest disability burden in the world. It is characterized by cognitive dysfunction, which has a significant impact on patients’ social function [1, 2]. Childhood adversity (CA) encompasses various types of physical or emotional abuse or neglect, as well as sexual abuse.Recent research indicates that CA is not only a significant risk factor for the onset and progression of MDD [3], but also strongly associated with diminished cognitive function in individuals with MDD [4]. For instance, Gould et al. pioneeredthe consideration of CA as a potential explanation for impaired cognitive function in MDD. They uncovered a link between early life adversity and cognitive deficits, specifically within a mixed cohort encompassing healthy participants, individuals diagnosed with depression, and patients suffering from PTSD [4]. According to a comprehensive review, CA is significantly linked to a deterioration in executive function, memory, and information processing speed in people with emotional disorders [5].

Diverse forms of CA may impact different aspects of cognitive performance, as evidenced by the inconsistent effects of CA on brain anatomy [6]. Neglect, not abuse, is related to impaired executive function in a study of teenagers [7]. Dannehl et al. [5] discovered that physical neglect and abuse are connected with poor language learning and executive skills, whereas emotional abuse may be paradoxically linked to improved executive functions. Grainger et al. found that neglect might have a more severe impact on cognitive function than abuse. Guo et al. also found that among untreated MDD patients, CA mostly impacted memory function, and the severity of adversityhad a dose-response relationship with memory performance. However, these investigations were typically limited to a single cognitive domain or had a small sample size [8]. and healthy control groups were rarely included to separate the specific effects of MDD [9]. Furthermore, investigations on different types of CA have found that neglect is related to reduced executive function, but the impact of abuse varies by subtype [10], implying the necessity for a multi-dimensional cognitive assessment to reveal the complex mechanisms of CA.

Some studies have found that adjusting for educational level reduces or even eliminates the impact of CA on cognitive function [10]. Educational level, as an essential measure of cognitive reserve, has been shown to mitigate the effects of some neurodegenerative disorders [11]. Other research has revealed that patients with a high level of education may be more susceptible to CA-related cognitive impairment due to their greater reliance on fluid intelligence [12]. This discrepancy suggests that sociodemographic characteristics (such as gender and educational level) may play a moderating role in the link between CA and cognitive function, but the evidence remains inconclusive [12].

Based on this, this study used a large sample of data (N = 2353) from the UK Biobank to systematically investigate: (1) the independent association between CA and multi-dimensional cognitive functions (fluid intelligence, reaction time, pairs matching, and prospective memory) in MDD patients; and (2) the moderating effects of gender and educational level on the aforementioned associations. By combining cumulative adversity scores and multi-dimensional neuropsychological evaluations, this study not only addresses previous study limitations such as small sample sizes and single evaluation indicators [8, 12], but also provides new evidence for investigating the heterogeneous mechanisms by which CA affects cognitive function. The study advances our understanding of the cognitive impairment process in MDD and provide a theoretical foundation for personalized management options for the group exposed to CA.

Methods

Study participants

This cross-sectional analysis utilized data from the UK Biobank, a large-scale cohort that recruited approximately 500,000 community-dwelling residents aged 40–69 years in the United Kingdom between 2006 and 2010. The cohort collected multi-dimensional data, including demographic characteristics, lifestyle factors, medical history, genetic information, and neuropsychological assessments. All participants provided written informed consent at baseline, and the study was approved by the Northwest Multi-Centre Research Ethics Committee in the UK .

A total of 8,222 individuals with major depressive disorder (MDD) were initially identified based on self-reported physician diagnoses. After excluding participants with baseline dementia (ICD-10 codes F00–F03), Parkinson’s disease (G20), or other severe neurological disorders (e.g., stroke, multiple sclerosis), as well as those with missing key variables (age, sex, education level) or incomplete childhood adversity data (n = 5,869), the final analysis included 2,353 participants (see flowchart in Fig. 1).

Fig. 1.

Screening process for inclusion of study participants

Childhood Adversity(CA) assessment

Childhood adversity was assessed using data from the 2016 online mental health questionnaire, employing the Childhood Trauma Screener (CTS), which consists of 5 items [13]. This is a cost-effective, validated, and relatively reliable screening tool for large-scale epidemiological studies. The five items represent physical neglect, emotional neglect, sexual abuse, physical abuse, and emotional abuse, respectively: (1) Someone to take to doctor when needed as a child (physical neglect); (2) Feel loved as a child (emotional neglect); (3)Sexually molested as a child (sexual abuse); (4) Physically abused by family as a child (physical abuse); (5) Felt hated by family member as a child (emotional abuse).Gan Yang et al.‘s article [14] was used to explain the scoring methodologies for each question. Each item is scored as 0 or 1. Possible responses include: never true, rarely true, sometimes true, often true, and very often true. For physical neglect, a score of 1 is assigned if the participant responded ‘never true’, ‘rarely true’, ‘sometimes true’, or ‘often true’; for emotional neglect, a score of 1 for ‘never true’, ‘rarely true’, or ‘sometimes true’; for sexual abuse, physical abuse, and emotional abuse, a score of 1 for ‘rarely true’, ‘sometimes true’, ‘often true’, or ‘very often true’. The summary score ranges from 0 to 5, with higher scores indicating greater exposure to childhood adversity.

Cognitive function assessment

Cognitive function was assessed in The UK Biobank using a touchscreen-based test battery [15]. Five cognitive tests were completed during the initial assessment period from 2006 to 2010. These tests are summarized as follows: (1)UKB pairs matching test (number of incorrect matches); (2) UKB reaction time test (mean time to correctly identify matches); (3) UKB fluid intelligence test (number of questions answered correctly in 2 min) (4) UKB prospective memory test (recall accuracy ) (5) UKB number memory test (the greatest number that was correctly recalled during the number memory test). Data for the Number Memory Test were available for only a subset of participants and were therefore not included in the current study. It is important to note that for the Pairs Matching Test, a higher number of incorrect matches indicates poorer cognitive performance, and for the Reaction Time Test, a longer average response time indicates poorer performance.

Covariate

A set of covariates was chosen based on previous research [14, 16], age, gender, educational level (higher, upper second, lower second and other), family income (<£18,000, £18,000 - £51,999, >£52,000), body mass index (BMI), Townsend Deprivation Index (TDI), and treatment status.

Statistical analysis

We described the participants’ basic characteristics by utilizing the mean (SD) of continuous variables and the count (%) of categorical variables. To investigate the relationship between CA and the cognition of depression patients, we used the generalized linear model (GLM) to assess the relationship between various CA (emotional neglect, physical abuse, emotional abuse, sexual abuse, and physical neglect) and cognitive functions (reaction time test, fluid intelligence test, pairs matching test, and prospective memory test). Continuous factors’ relationships were represented by the β and 95% confidence interval (95% CI), whereas categorical variables were represented by the odds ratio (OR) (95% CI). This study used two models. Model 1 was adjusted for gender and age. Model 2 was based on Model 1 but has been adjusted for educational level, family income, BMI, TDI, and whether therapy was sought.

Given that Model 2 incorporated a broader set of covariates extensively validated in the literature as being closely associated with both childhood adversity (CA) exposure and cognitive function outcomes, we stratified the sample by sex (male/female) and education level (higher/upper secondary/lower secondary/other) within Model 2. This allowed us to examine the associations between childhood adversity and cognitive function separately across different gender and educational backgrounds. All analyses were carried out with version 4.1.1 of the R Statistical program (R Project for Statistical Computing) and SPSS 26. P < 0.05 indicates significance.

Results

Demographic characteristics of all the participants

Table 1 presents the baseline characteristics of 2,353 participants. The participants’ average age was 52.4 (7.3) years old. A total of 1,596 participants (67.8%) were famale. The pairs matching test yielded a median of three erroneous pairings. The median time to complete the response time test was 520 microseconds. The median score on the fluid intelligence exam was seven points. Furthermore, 98.4% of participants properly answered the prospective memory test.

Table 1.

Participants’ general characteristics

Characteristic
Overall number of participants	2353
Age (Mean ± SD)	52.4 ± 7.3
Sex (%)
Female	1596(67.8)
Male	757(32.2)
Townsend deprivation index (Mean ± SD)	-1.4 ± 3.0
Body mass index(Mean ± SD)	26.6 ± 4.7
Educational levels (%)
Higher	1165(49.5)
Upper second	570(24.2)
Lower second	411(17.5)
Other	207(8.8)
Income (%)
<£18 000	443(18.8)
£18 000–£51 999	1243(52.8)
>£52 000	667(28.3)
Whether therapy was sought (%)
Yes	1297(55.1)
No	1056(44.9)
UKB pairs matching test (median [IQR])	3[2,5]
UKB reaction time test (median [IQR])	520[469,590 ]
UKB prospective memory test (%)
Pass	798(98.4)
No	13(1.6)
UKB fluid intelligence test (median [IQR])	7[5,8]
Cumulative childhood adversity score (0–5)(%)
0	934(39.7)
1	611(26.0)
2	390(16.6)
3	239(10.2)
4	141(6.0)
5	38(1.6)

Data were presented as number of participants (percentage), mean ± SD (normal distribution), or median interquartile range [IQR] (abnormal distribution)

UKB UK Biobank

Association between childhood adversity and cognitive function in MDD

Table 2 shows the relationships between individual as well as cumulative CA, and cognitive test. Increased ratings for physical abuse, physical neglect, and cumulative CA are significantly associated with poorer fluid intelligence. Increased ratings for emotional abuse and physical neglect are connected with poorer cognitive performance in pairs matching test (P < 0.05). The above associations are stable in Models 1 and 2. An elevated physical neglect score was associated with poorer performance on the reaction time test in Model 1. No significant association was found between childhood maltreatment and performance on the prospective memory test (Table 3, 4, 5).

Table 2.

Associations of childhood adversity with cognitive function in major depressive disorder. Associations of childhood adversity with reaction time test

	Model 1		Model 2
Childhood adversity	β (95% CI)	P value	β (95% CI)	P value
Emotional neglect	3.214 (-5.206,11.635)	0.454	1.210 (-7.327,9.747)	0.781
Physical abuse	6.854 (-2.679, 16.387)	0.159	4.232 (-5.429,13.893)	0.391
Emotional abuse	4.389 (-4.7402, 13.519)	0.346	2.024 (-7.258,11.305)	0.669
Sexual abuse	2.726 (-9.522, 14.975)	0.663	0.191 (-12.147,12.529)	0.976
Physical neglect	11.649 (0.959, 22.338)	0.033	9.083 (-1.689, 19.856)	0.098
Cumulative childhood adversity score	2.872 (-0.256, 5.999)	0.072	1.701 (-1.510, 4.912)	0.299

Table 3.

Associations of childhood adversity with cognitive function in major depressive disorder. Associations of childhood adversity with fluid intelligence test

	Model 1		Model 2
Childhood adversity	β (95% CI)	P value	β (95% CI)	P value
Emotional neglect	-0.281 (-0.574,0.012)	0.061	-0.261 (-0.543,0.021)	0.070
Physical abuse	-0.434 (-0.757,-0.111)	0.009	-0.420 (-0.729,0.111)	0.008
Emotional abuse	-0.161 (-0.478, 0.157)	0.321	-0.178 (-0.485,0.128)	0.254
Sexual abuse	-0.137 (-0.554 0.280)	0.520	-0.198 (-0.598,0.202)	0.332
Physical neglect	-0.779 (-1.138 -0.421)	2.3e-05	-0.578 (-0.922,-0.235)	0.001
Cumulative childhood adversity score	-0.180 (-0.286 -0.073)	0.001	-0.169 (-0.272, -0.066)	0.001

Table 4.

Associations of childhood adversity with cognitive function in major depressive disorder. Associations of childhood adversity with pairs matching test

	Model 1		Model 2
Childhood adversity	β (95% CI)	P value	β (95% CI)	P value
Emotional neglect	0.165 (-0.076 0.405)	0.179	0.107 (-0.136,0.351)	0.388
Physical abuse	0.058 (-0.214 0.330)	0.674	0.031 (-0.244, 0.307)	0.824
Emotional abuse	0.334 (0.073 0.594)	0.012	0.285 (0.020, 0.549)	0.035
Sexual abuse	-0.021 (-0.370,0.329)	0.907	-0.073 (-0.426, 0.279)	0.682
Physical neglect	0.380 (0.075 0.685)	0.015	0.328 (0.020, 0.636)	0.037
Cumulative childhood adversity score	0.099 (0.010 0.189)	0.029	0.077 (-0.015, 0.168)	0.101

Table 5.

Associations of childhood adversity with cognitive function in major depressive disorder. Associations of childhood adversity with prospective memory test

	Model 1		Model 2
Childhood adversity	OR (95% CI)	P value	OR (95% CI)	P value
Emotional neglect	1.32 (0.44-4.00)	0.620	1.32 (0.41–4.25)	0.643
Physical abuse	1.10 (0.33–3.64)	0.872	0.94 (0.27–3.31)	0.923
Emotional abuse	0.83 (0.25–2.77)	0.756	0.75 (0.21–2.64)	0.651
Sexual abuse	1.85 (0.49–6.98)	0.362	1.59 (0.39–6.43)	0.513
Physical neglect	1.97 (0.59–6.52)	0.269	1.69 (0.47–5.99)	0.419
Cumulative childhood adversity score	1.14 (0.78–1.66)	0.500	1.08 (0.73–1.59)	0.707

Model 1 was adjusted for sex and age

Model 2 was further adjusted for educational levels, Townsend Deprivation Index, Body mass index , Income and Have ever sought treatment based on model 1

Association between CA and cognitive function in different genders

Supplement Table 1, stratified by gender and including all covariates (Model 2), revealed that among women, with the exception of no association between sexual abuse and fluid intelligence(β = 0.083, 95% CI=-0.858 to 1.024, P = 0.863), all other forms of CA and cumulative CA increased scores were significantly associated with poorer cognitive performance in fluid intelligence. An increase in female emotional abuse(β = 0.398, 95% CI = 0.086 to 0.711, P = 0.013), physical neglect(β = 0.491, 95% CI = 0.130 to 0.853, P = 0.008), and cumulative abuse scores(β = 0.125, 95% CI = 0.018 to 0.23, P = 0.022) is associated with poorer cognitive performance on the pairs matching test. However, there was no link between CA and cognitive function in men.

Association between CA and cognitive function in different educational levels

After stratifying by educational level and adding all covariates in Supplement Table 2 (Model 2), the results revealed that among patients to a higher educational level, an increase in the physical neglect score(β = 17.785, 95% CI = 1.641 to 33.929, P = 0.031) was related with poorer cognitive performance in the reaction time test. Increased scores for emotional neglect(β=-0.496, 95% CI=-0.920 to -0.073, P = 0.022), physical abuse(β=-0.674, 95% CI=-1.143 to -0.204, P = 0.005), physical neglect(β=-0.593, 95% CI=-1.165 to -0.020, P = 0.043), and cumulative abuse(β=-0.254, 95% CI=-0.409 to -0.099, P = 0.001) were connected with lower cognitive function in fluid intelligence. In the pairs matching test, patients with an educational level of Other had poorer cognitive performance scores for emotional neglect(β = 0.989, 95% CI = 0.129 to 1.849, P = 0.025), emotional abuse(β = 1.206, 95% CI = 0.243 to 2.169, P = 0.015), physical neglect(β = 1.677, 95% CI = 0.723 to 2.631, P = 0.001), and cumulative abuse(β = 0.490, 95% CI = 0.181 to 0.798, P = 0.002).

Discussion

This study, based on UK Biobank data, systematically investigated the associations between different types of childhood adversity (CA) and multidimensional cognitive function in patients with major depressive disorder (MDD), further revealing the moderating roles of sex and education level. The cognitive assessments covered a range from higher-order abstract reasoning (fluid intelligence test) to basic information processing (reaction time and pairs matching tests). The impact of CA on cognitive function was not uniform or homogeneous but exhibited a complex pattern characterized by adversity-type specificity, cognitive-domain specificity, and sociodemographic specificity.

Our findings align with previous studies indicating that CA negatively affects cognitive function [5, 17]. Increased severity of physical abuse, physical neglect, and cumulative adversity scores were significantly associated with poorer performance in fluid intelligence tests. Fluid intelligence represents the ability to engage in abstract reasoning, problem-solving, and cognitive flexibility in novel situations, primarily relying on the prefrontal cortex and its coordination with parietal networks. This supports the view from prior research that “childhood adversity impairs higher-order cognitive functions through chronic stress and neurobiological alterations"²¹. Specifically, severe stress experiences during childhood (particularly threat and deprivation) may lead to persistent dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis, enhanced neuroinflammation, and impaired synaptic plasticity, ultimately causing structural and functional damage in advanced brain regions such as the prefrontal cortex, thereby undermining fluid intelligence [18–20]. Neuroimaging studies demonstrating that CA exposure is closely linked to reduced prefrontal cortex volume and aberrant functional connectivity in adults further corroborate our findings [19].

Emotional abuse and physical neglect were associated with increased errors in pairs matching tests (reflected by a higher number of incorrect pairs). Pairs matching tests primarily assess visuospatial working memory and basic information processing efficiency, representing more fundamental cognitive processes. This suggests that certain types of adversity affect not only abstract reasoning but also more basic cognitive components, echoing Gould et al., findings on the broad cognitive impairments associated with childhood maltreatment [4]. Notably, physical neglect was linked to both fluid intelligence decline and pair matching deficits, indicating its widespread rather than localized negative impact on cognitive function. The association of physical neglect with both higher-order abstract reasoning and basic information processing supports the notion that “deprivation-type” adversity has more profound effects on cognition [21], consistent with previous conclusions that neglect may exert longer-lasting cognitive damage compared to abuse [7, 22].

In reaction time tests, significant associations were observed only in certain subgroups, manifested as prolonged reaction times (reduced information processing speed). This further indicates that different cognitive metrics vary in their sensitivity to childhood adversity, and a single cognitive test is insufficient to fully capture the cognitive consequences of CA, aligning with the emphasis on cognitive domain heterogeneity in MDD patients by Rock et al. [2].

Stratified analyses revealed that the negative associations between CA and cognitive function were almost entirely concentrated in female MDD patients. Among women, multiple adversity types were associated with fluid intelligence decline, and emotional abuse and physical neglect were also linked to poorer pair matching performance. In contrast, no significant associations were found in male patients. This result supports prior evidence that females are more vulnerable to the long-term psychological and neurobiological consequences of childhood adversity (particularly emotional abuse and neglect), potentially due to sex-specific neuroendocrine responses, differences in emotion regulation strategies, and gender variations in the development and plasticity of prefrontal-limbic systems [23, 24]. This finding suggests that ignoring sex differences in MDD populations may obscure important CA-cognitive impairment linkages and provides a basis for more targeted cognitive assessments and interventions in female patients.

Education level, as a key indicator of cognitive reserve, exhibited complex moderating patterns in the CA-cognitive function relationship. In patients with higher education levels, multiple adversities were associated with fluid intelligence decline, and physical neglect was also linked to poorer reaction time test performance (prolonged reaction times). Reaction time tests measure basic information processing speed and attentional vigilance. This contrasts with some studies suggesting that higher education acts as a protective factor for cognitive function in MDD [11], indicating that greater cognitive reserve may not fully shield individuals from CA-induced damage to both higher-order abstract reasoning and basic information processing [25]. Highly educated individuals often rely more deeply on prefrontal cortex-mediated abstract reasoning when solving problems; when childhood adversity impairs the neural substrates of these advanced functions, this dependency may make deficits more apparent [26]. In patients with lower education levels, multiple adversities were significantly associated with poorer pair matching performance, while links to fluid intelligence were relatively weaker. This reflects a domain separation in cognition across educational backgrounds: individuals with lower education may have limited cognitive reserve resources, making their cognitive function more directly and earlier impacted by CA at basic information processing stages. The protective effect of cognitive reserve is not universally applicable but has clear boundaries, offering new explanations for inconsistent findings on education’s role in previous studies and resonating with conclusions from Barczyk et al.‘s systematic review in mood disorders [12].

This study did not observe a significant association between CA and prospective memory tests. This result may reflect the true unassociation, or it may be related to the insufficient sensitivity of the UK Biobank prospective memory task itself, which precisely corresponds to the previous assessment results of the validity of the UK Biobank cognitive test by Fawnes-Ritchie and Deary [15].

This study systematically unveils the heterogeneity in how childhood adversity affects multidimensional cognitive function in MDD patients, emphasizing that clinical practice should integrate systematic CA screening and assessments covering different cognitive levels (e.g., abstract reasoning, processing speed) to precisely identify patients’ specific cognitive impairment patterns [13]. Cognitive intervention should be individualized and domain-targeted, as patterns of cognitive damage vary by adversity type, sex, and education background. For example, patients with a history of physical neglect and higher education may require simultaneous focus on abstract reasoning and basic information processing training, whereas those with lower education exposed to emotional abuse might need interventions targeting basic information processing.Understanding the moderating roles of sex and education level aids in setting realistic treatment expectations and selecting optimal intervention strategies for different patient subgroups [27].

This study is the first to leverage the large-scale UK Biobank cohort (N = 2353), integrating retrospective data on childhood adversity with multidimensional cognitive assessments (encompassing fluid intelligence, reaction time, pair matching, and prospective memory). A cumulative adversity score (ranging from 0 to 5) was employed to quantify the cumulative effect of multiple adverse experiences, addressing limitations of previous studies such as small sample sizes and restricted assessment metrics. Furthermore, this study has several limitations: the cross-sectional design makes it difficult to clarify the causal inferences regarding the relationship between childhood adversity and cognitive impairment; the cognitive test battery did not comprehensively cover complex executive functions [15]; and retrospective self-report data may introduce recall bias. Particularly, individuals with major depressive disorder might exhibit systematic recall biases influenced by their current mood state (e.g., depressive affect), potentially amplifying the observed associations between adversity and negative cognitive outcomes. However, prior research indicates that adult retrospective reports of childhood trauma demonstrate long-term stability and are associated with objective indicators [28]. In this study, we utilized a standardized childhood trauma screening tool (CTS) and focused on relatively explicit and specific behavioral events (e.g., “experiencing physical abuse,” “not being taken to a doctor when needed”), which helped enhance the credibility of the reports. Future research should adopt multimodal approaches, as advocated by Lupien et al., integrating longitudinal tracking, neuroimaging (e.g., hippocampal volume, prefrontal-limbic system connectivity), and epigenetic markers (e.g., FKBP5 methylation) to delineate the dynamic pathways through which childhood adversity influences cognition [14, 29].

Conclusion

In summary, this study highlights the complexity of the relationship between CA and cognitive function in MDD patients, underscoring the importance of multidimensional assessment and personalized interventions. It provides key evidence for future mechanistic exploration and clinical practice.

Authors’ contributions

Yang, Sun and Cai designed and supervised this study. Li, Hao, Ren and Shi were responsible for data analysis and manuscript drafting. Zhang and Yang revised the manuscript. Li, Hao and Ren carried out the experimental procedures. All authors reviewed and approved the final manuscript.

Funding

This work was supported by the Health Commission scientific research of Shanxi Province(2023037, 2024049), National Clinical Key Specialty Construction Project (Project No.2025-ZZ-003), National Natural Science Foundation of China (82171534), Natural Science Foundation of Shanxi Province (20210302123251, 20210302123257), Natural Science Foundation of Shanxi Province for youth (20210302124193), 136 Medical Rejuvenation Project of Shanxi Province, Four “batches” innovation project of invigorating medical through science (2023RC004) and technology of shanxi province and the special fund for Science and Technology Innovation Teams of Shanxi Province (202204051001027).

Data availability

Publicly available datasets were analyzed in this study. This data can be found here: this study used the UK Biobank resource with the application ID 94166 and 211475. Researchers can access the UK Biobank by applying to the UK Biobank official website ( https://www.ukbiobank.ac.uk/).

Declarations

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.